babl is a dynamic, any to any, pixel format translation library.
It allows converting between different methods of storing pixels known as pixel formats that have with different bitdepths and other data representations, color models and component permutations.
A vocabulary to formulate new pixel formats from existing primitives is provided as well as the framework to add new color models and data types.
Features
- Fast.
- Accurate.
- Stable, small API.
- Self profiling and optimizing.
- ANSI C, works on win32, linux and mac, 32bit and 64bit systems.
- Extendable with new formats, color models, components and datatypes.
- Reference 64bit floating point conversions for datatypes and color models.
GEGL through GeglBuffer provides tiled buffers with on disk storage as well as linear buffers with accessor functions for efficient data access transparently using babl fishes for translation to the desired pixel formats.
Download
The latest versioned development version of babl can be found in ftp://ftp.gtk.org/pub/babl/.
Babl uses git. The main repository is hosted by GNOME. It can be browsed online and cloned with:
git clone git://git.gnome.org/babl
The following is a list of the major changes that have gone into each babl release. If there are significant improvements to babl when a GEGL release is done a babl release is most often put out just prior to the GEGL release.
- 2012-03-30 babl-0.1.8
- LUT based speedups for gamma correction / uncorrection.
- 2012-03-30 babl-0.1.8
- Added support for indexed/pallette based formats, constified API.
- 2011-11-18 babl-0.1.6
- Build improvements, remove blatantly wrong conversions from extensions, made it possible to distinguish format_n formats from others, improvements to vala/gobject introspection support.
- 2011-01-20 babl-0.1.4
- Improved cross platform build ability, added code to handle n-component formats, remove extraenous runtime profiling that impacted performance.
- 2010-01-15 babl-0.1.2
- Made babl mostly threadsafe (sufficient for GIMP/GEGLs needs). Streamlined core functionality, reducing number of function calls and removing overeager instrumentation.
- 2009-05-20 babl-0.1.0
- Support for RGBA representation in the frequency domain, un-pre-multiply close-to-zero alpha values to complete black instead of slightly brighter than black, add a BABL_ALPHA_THRESHOLD constant to the API, do a complete overhaul of the babl API and do some changes to increase portability of the library.
- 2008-06-12 babl-0.0.22
- Speed improvements due to: coalesced hashing, early bail out upon creation of duplicate formats, caching of non existing paths. Large amounts of the core has also been reviewed and refactored. Improved conversion coverage in the matrix.
- 2008-02-27 babl-0.0.20
- Builds on OSX Check <0.0 and >1.0 values conversions for accuracy as well.
Documentation
When using BablFishes to do your conversions, you request a fish to convert between two formats, and an optimal fish to babls capability is provided that you can use to do your conversions. Babl also provides the capability to describe new formats based on a vocabulary of user registered color models and data types.
Babl provides a base vocabulary in BablBase and some extensions that are thought to be generally useful.
When performing further extensions to the vocabulary of babl, the internal consistency is governed by reference conversions that operate on double (64 bit floating point values). The only color model created during BablCore bootstrap is RGBA (linear light RGB, 0.0 - 1.0, with a linear 0.0 - 1.0 opacity channel) backed by the double datatype. Defined similarily to scRGB using 64bit floating point.
If babls conversion isn't fast enough, you can provide your own conversion shortcut between two formats. The registered shortcut might also be used by babl as an intermediate conversion when constructing BablFishes for other conversions.
Babl extensions are shared objects. If you have already developed some fast conversion functions, wrapping them as babl extensions should not take much time and will speed up babl for other users as well.
Usage
babl_process (babl_fish (source_format, destination_format), source_buffer, destination_buffer, pixel_count);
The processing operation that babl performs is copying including conversions if needed between linear buffers containing the same count of pixels, with different pixel formats.
int width = 123, height = 581, pixel_count = width * height; const Babl *srgb = babl_format ("R'G'B' u8"); const Babl *lab = babl_format ("CIE Lab float"); Babl *rgb_to_lab_fish = babl_fish (srgb, lab); float *lab_buffer; unsigned char *srgb_buffer; babl_init (); srgb_buffer = malloc (pixel_count * babl_format_get_bytes_per_pixel (srgb)); lab_buffer = malloc (pixel_count * 3 * sizeof (float)); ...... load data into srgb_buffer ....... babl_process (rgb_to_lab_fish, srgb_buffer, lab_buffer, pixel_count); ...... do operation in lab space ........ babl_process (babl_fish(lab, srgb), lab_buffer, srgb_buffer, pixel_count); /* the data has now been transformed back to srgb data */
If the existing pixel formats are not sufficient for your conversion needs, new ones can be created on the fly. The constructor will provide the prior created one if duplicates are registered.
const Babl *format = babl_format_new (babl_model ("R'G'B'"), babl_type ("u16"), babl_component ("B'"), babl_component ("G'"), babl_component ("R'"), NULL);
Vocabulary
+-
Shortcut Coverage
The diagram shown below visualizes the coverage of current shortcut conversions. Dots indicate a direct conversion is provided for, the height of the bar indicates the number of conversions steps needed in a chain of conversions. A DHTML version is also available.
▂· ·▁·· ··· ▁ ▁ ▂ ▁▁ ▁ · ···· ── 0 RGBA double
── 1 RGB double
── 2 RaGaBaA double
── 3 R'G'B' double
· ▁ ▁▂▁▁ ▁▁▁ · ▂ ▁ ▂▂ ▂ ▁ ▁▁▁▁ ── 4 R'G'B'A double
· ▁▂▁▁ ▁▁▁ ▂ ▃ ▂▂ ▂ ▁ ▁▁▁ ── 5 R'aG'aB'aA double
▁ ▂ ···▂▂ ▂▂▂ ▁ ▁ ·· ▁· ▂ ▁ ▂ ▃▃▃▃ ── 6 R'G'B' u8
▁ ▁▂ ·▁▃▂ ▃▂▂ · ▁ ▁ ▁▁ ▂ ▂ ▂▂▂▂ ── 7 R'G'B'A u8
· ▁▁ ·▁▁ ▁▁▁ ▂ · ▃ ·· · ▁ ▁▁▁▁ ── 8 RGBA float
▁ ▃▂ ▂▂ ▂▂▂ ▁ ▁ · ▂· ▂ ▁ ▂ ▃▂▃▂ ── 9 RGB float
· ▁▁ ▁ ▁ ▁▁▁ ▂ ▃ ▃ ▃▂ ▃ ▂ ▁ ▁▁▁▁ ──10 Y double
· ▁ ▁▂ ▁▁▁ ▂ ▂ ▃ ▂▂ ▂ ▁ ▁▁▁▁ ──11 YA double
──12 YaA double
──13 Y' double
──14 Y'A double
· ▁▁ ▁▂▁▁ ▁▁ ▂ ▃ ▂▂ ▂ ▁ ▁▁▁ ──15 Y'aA double
· ▁▁ ▁▂▁▁ ▁ ▂ ▃ ▃ ▃▂ ▃ ▂ ▁ ▁▁ ▁ ──16 Y'CbCr double
· ▁▁ ▁▂▁▁ ▁ ▂ ▂ ▃ ▂▂ ▂ ▁ ▁▁▁▁ ──17 Y'CbCrA double
──18 Y'CbCr u8
▁ ·▂ ·▁▂▂ ▂▂▂ ▁ · ▁▁ ▁ ▂ ▂▂▂▂ ──19 R'G'B'A float
▂ ▁▃ ▁▂▃▃ ▃▃▃ ▁ ▁ ▂▂ ▂ ▃ ▃▃▃▃ ──20 R'G'B'A u16
▂ ▂ ▃▃▃▃ ──21 R'aG'aB'aA float
▃ ▂ ▂▃ ▃▃▃ ▁ ▂ ▃▃ ▃ ▃ ▃▃▃▃ ──22 R'aG'aB'aA u16
▃ ▂ ▂▃ ▁ · ▂ ▃▃ ▃ ──23 R'aG'aB'aA u8
▂ ▁▃ ▁·▃▃ ▃▃▃ · · ▃▁ ▃ ▂ ▃ ▃▃▃▃ ──24 R'G'B' float
▃ ▂ ▂▁ ▁ ▁ ▃▂ ▃ ──25 R'G'B' u16
▁ · ▁ ▁ ·· ▂ ▁ ──26 Y'A float
▂ ▁ · ▂ ▁ ▁ · ──27 Y'aA float
▂ ▁ ▁ · · ──28 Y' float
▂ ▁ ▂ ▂ ·▁ ▃ ▂ ──29 Y'A u16
▃ ▂ ▁ ▃ ▁·▂ ▁ ──30 Y'aA u16
▃ ▂ ▂ ▁ ▁▂· ──31 Y' u16
▁ ▂▂ ·▂▂▃ ▂▂▃ ▁ ▂ ▂ ·▁▁ ▃ ▂ ▁▂ ▁·▁▁ ▃ ▂▂▂▂ ──32 Y'A u8
▃ ▂ ▁ ▃ ▁·▂ · ──33 Y'aA u8
▁ ▃▂ ·▁▂▂ ▂▂▂ ▂ ▂ ▁ ▁▁· ▂▁ ▂ ▁▁▁· ▂ ▂▂▂▂ ──34 Y' u8
▃ ▂ ▂▁ ▁ ▁ · ▃▂ ▃ ──35 Y'CbCr float
▂ ▁▃ ▁▂▃▃ ▃▃ · ▂ ▁ ▂▂ ▂ ▃ ▃▃▃▃ ──36 Y'CbCrA float
▁ ▂▂ ·▁▂▂ ▂▂▂ ▁ ▂ ▂ ▁▁ ▁ ▂ ▂▂▂▂ ──37 RGBA u16
▁ ▂▂ ·▁▂▂ ▂▂▂ ▃ ▂ ▂ · · ▂ ▃▂▂▂ ──38 RGBA u8
▁ ▂▂ ·▁▂▂ ▂▂▂ ▃ ▁ ▂ ▁ ▂ ▂▂▂▂ ──39 RaGaBaA float
▂ ▃▃ ▁▂▃▃ ▃▃▃ ▂ ▁ ▃ ▂ ▂ ▃ ▃▃▃▃ ──40 RaGaBaA u16
▂ ▃▃ ▁▂▃▃ ▃▃▃ ▂ ▂ ▃ · ▃ ▃▃▃▃ ──41 RaGaBaA u8
▂ ▃▃ ▁·▃▃ ▃▃▃ ▂ ▂ ▁ ▃▁ ▂ ▃ ▃▃▃▃ ──42 RGB u16
▁ ▃▃ ··▃▂ ▂▂▂ ▂ ▃ ▃ ·· · ▂ ▃▃▂▃ ──43 RGB u8
▁ ▂▂ ·▁▂ ▂▂▂ ▃ ▁ ▂ ▁▁ ·· ▂ ▂▂▂▂ ──44 YA float
▂ ▃▃ ▁▂▃▃ ▃▃▃ ▂ ▂ ▃ ▂· ▂ ▃ ▃▃▃▃ ──45 YaA float
▂ ▃▃ ▁· ▃ ▃▃▃ ▂ ▂ ▁ ▃▁ ▃ ▂· ▃ ▃▃▃▃ ──46 Y float
▂ ▃▃ ▁▂▃▃ ▃▃▃ ▂ ▂ ▃ ▂▂ ▂·▁ ▃ ▃▃▃▃ ──47 YA u16
▃ ▂▃ ▃ ▂ ▃▁ ▃ · ──48 YaA u16
▃ ▂▁ ▃ ▃ ▂ ▃▂ ▃▁▁· ──49 Y u16
▁ ▂▃ ·▁▂▂ ▂▂▃ ▁ ▃ ▃ ▁▁ ▁·▁▁ ▂ ▃▂▂▂ ──50 YA u8
▃ ▂▃ ▃ ▂ ▁ ▁·▂ ──51 YaA u8
▁ ▂ ·▁ ▂ ▂▂▂ ▃ ▃ ▂ ▂▂ ▂ ▁▁▁· ▂ ▂▂▂▂ ──52 Y u8
──53 B'aG'aR'aA u8
──54 cairo-ARGB32
──55 cairo-RGB24
──56 cairo-A8
· ▁▁ ▁▂▁▁ ▁▁▁ ▂ ▃ ▃ ▃▂ ▃ ▂ ▁▁▁▁ ──57 CMYK double
──58 CMYK float
· ▁▁ ▁▂▁▁ ▁▁▁ ▂ ▃ ▃ ▂▂ ▃ ▂ ▁ ▁▁▁ ──59 CIE Lab double
· ▁▁ ▁▂▁▁ ▁▁▁ ▂ ▃ ▂▂ ▂ ▁ ▁▁ ──60 CIE Lab alpha double
· ▁▁ ▁▂▁▁ ▁▁▁ ▂ ▃ ▃ ▃▂ ▃ ▂ ▁ ▁ ──61 CIE LCH(ab) double
· ▁▁ ▁▂ ▁ ▁▁▁ ▂ ▃ ▂▂ ▂ ▁ ▁▁ ──62 CIE LCH(ab) alpha double
──63 CIE Lab float
──64 CIE Lab alpha float
──65 CIE Lab u8
──66 CIE Lab u16
──67 CIE LCH(ab) float
──68 CIE LCH(ab) alpha float
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||11111111112222222222333333333344444444445555555555666666666
012345678901234567890123456789012345678901234567890123456789012345678
total length: 2387
total cost : 796267
Environment
If the environment variable BABL_STATS is set containting a html pixel format conversion usage matrix will be written to /tmp/babl-stats.html. This allows figuring out which conversions is taking up time during processing, and what shortcuts extensions might be created or improved to make babl do it's job faster.
Through the environment variable BABL_TOLERANCE you can control a speed/performance trade off that by default is set very low (0.000001) values in the range 0.01-0.1 can provide reasonable preview performance by allowing lower numerical accuracy
.Extending
For samples of how the current internal API specification of data types, color models, and conversions look in the extensions/ directory. The tables in this HTML file is directly generated based on the data registered by BablCore (double and RGBA), BablBase (core datatypes, and RGB models), extensions (CIE Lab, naive CMYK, various shortcut conversions).
Directory Overview
babl-dist-root
│
├──babl the babl core
│ └──base reference implementations for RGB and Grayscale Color Models,
│ 8bit 16bit, and 32bit and 64bit floating point.
├──extensions CIE-Lab color model as well as a naive-CMYK color model.
│ also contains a random cribbage of old conversion optimized
│ code from gggl. Finding more exsisting conversions in third
│ part libraries (hermes, lcms?, liboil?) could improve the
│ speed of babl.
├──tests tests used to keep babl sane during development.
└──docs Documentation/webpage for babl (the document you are reading
originated there.
TODO
- Support for conversions between formats that have higher dimensionality than RGBA. (arbitrary meaningless components)
- Support for datatypes that are not a multiple of 8bit.
- dithering
- dynamic data attached to format, or internal handling of icc through a registered set of profiles
- Stocastic runtime profiling with running average of the peformance for conversions, to avoid problems incurred if other cpu intense is happening when babl tests a conversions.
Copyright
Babl is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.
Authors
- Øyvind Kolås pippin at gimp.org
- Original author.
- Sven Neumann sven at gimp.org
- Build sanity and optimizations.
- Michael Natterer mitch at gimp.org
- Build sanity.
- Kevin Cozens kcozens at cvs.gnome.org
- Build sanity.
- Tim Mooney
- Portability fixes.
- Michael Schumacher schumaml at cvs.gnome.org
- win32 support for dynamic extensions.
- Portability fixes.
- Jan Heller jheller at svn.gnome.org
- Optimizations, refactoring and documentation.
- Mukund Sivaraman muks at mukund.org
- Sparse fixes and sanity.
- dmacks at netspace.org
- Build sanity
- Sam Hocevar
- Build sanity.
- Zhang Junbo
- Frequency domain color model.
- Martin Nordholts
- Optimizations and API overhaul.
- Gary V. Vaughan
- Multiplatform build support.
- Stanislav Brabec
- Portability patch.
- Hubert Figuiere
- C++ friendliness patch.
- Danny Robson
- Code cleanups.
- Fryderyk Dziarmagowski freetz at gmx.net
- Code cleanups.
- Daniel Paredes García danipga at gmail.com
- Initial work to make babl threadsafe
- Rupert Weber gimp at leguanease.org
- Documentation and other improvements
